Pricing Misinformation

Evidence from Housing Transactions

Alexander Cardazzi

Old Dominion University

Crocker H. Liu

Cornell University

Adam D. Nowak

West Virginia University

Patrick S. Smith

University of North Carolina at Charlotte

Introduction

Economists often estimate WTP for urban amenities by using changes in housing prices. School quality and housing is an often studied relationship.

Using data from the Chicago public schools, we estimate that serious cases of teacher or administrator cheating on standardized tests occur in a minimum of 4–5 percent of elementary school classrooms annually. (Jacob and Levitt 2003)

How does grade inflation or, more explicitly, cheating factor into this? To explore this, we study a late 2000s cheating scandal in Atlanta.

APS Background

Criterion-Referenced Competency Tests (CRCT) has three pass levels:

Does not meet expectations
Meets expectations
Exceeds expectations

Atlanta Public Schools Cheating Scandal

Teachers and administrators were placed under extreme pressure to meet certain CRCT standards. Meeting standards meant bonuses while failing to do so meant termination.

178 Educators implicated | 35 indicted | 44/56 schools cheated on the 2009 CRCT

Defining Cheating

To measure cheating, the state employed an “erasure” analysis. Calculate the \(z\)-Score of the number of wrong-to-right erasures in a grade-subject-class; \(z > 3\) indicates the classroom is cheating.

	Georgia	APS
Clear of Concern	80%	23%
Minimal Concern	10%	8%
Moderate Concern	6%	18%
Severe Concern	4%	51%

APS accounts for over half of the 4% of severe cases.

Cheating Severity

Erasures vs CRCT Pass Rates

Erasures vs Exceed Rates

Erasures vs Meets Rates

APS vs Georgia

Specifying a Regression

Most studies estimate:

\[log(\text{Price}_{ist}) = X_{it}\beta + \gamma_1\widehat{\text{Pass}}_{st}+ \epsilon_{ist}\]

but we know:: \(\widehat{\text{Pass}}_{st} = \text{Pass}_{st} + \text{Inflation}_{st}\); \(cov(\text{Pass}_{st}, \text{Inflation}_{st}) < 0\)

We can only observe \(\widehat{\text{Pass}}_{st}\) and \(\text{Erasures}_{s,t\in2009, 2010}\)

\[log(\text{Price}_{ist}) = X_{it}\beta + f(\widehat{\text{Pass}}_{st}, \text{Erasures}_{s,t=2009}) + \epsilon_{ist}\]

We control for square footage, and include FEs for bedrooms-by-bathrooms, 0.5-mile grid cells, school year, and indicators for distressed transactions (REO / short sales).

Grade Inflation and Housing Prices
	(1)	(2)	(3)	(4)	(5)
Pass Rate	0.216		0.157	-0.336*	0.424
	(0.161)		(0.175)	(0.164)	(0.302)
Pass Rate x Post	0.894***		0.700***	0.763***	0.570***
	(0.109)		(0.109)	(0.104)	(0.158)
'09 Erasures		-0.205	-0.282*
		(0.125)	(0.127)
'09 Erasures x Post		-0.478***	-0.187**	-0.305***	-0.566
		(0.045)	(0.057)	(0.058)	(0.474)
Pass Rate x '09 Erasures					-1.349*
					(0.660)
Pass Rate x '09 Erasures x Post					0.180
					(0.582)
Num.Obs.	15535	15535	15535	15535	15535
School FEs	N	N	N	Y	Y
R2	0.857	0.857	0.858	0.862	0.862

Results

Next Steps

When did erasures start / stop?
How to map erasures to inflation?
Interpretation?

Bibliography

Jacob, Brian A, and Steven D Levitt. 2003. “Rotten Apples: An Investigation of the Prevalence and Predictors of Teacher Cheating.” The Quarterly Journal of Economics 118 (3): 843–77.